A method and device for clustering the scattering path of a wireless channel

Abstract

The invention provides a wireless channel scattering path clustering method and device. The method comprises: acquiring scattering path parameters corresponding to all scattering path components in awireless channel; calculating the relative density of each scattering path component according to the scattering path parameters; selecting a scattering path component of which the relative density isequal to a preset density value as a scattering path cluster center; defining a high-density field of each scattering path component, establishing a directed path diagram, and dividing the scatteringpath components communicated with the same scattering path cluster center in the directed path diagram into the same scattering path cluster; establishing an undirected path diagram for each scattering path cluster center; and for the divided scattering path clusters, combining the scattering path clusters corresponding to two scattering path cluster centers which are communicated with each otherin the undirected path graph and have the relative density of the scattering path components on any communication path greater than a preset density threshold to obtain a final clustering result. According to the method, the characteristic change of the scattering path component can be distinguished more accurately without cluster prior information, and the method can be used for channel modelingin the field of wireless communication.

Description

technical field [0001] The invention relates to the technical field of wireless communication, in particular to a wireless channel scattering path clustering method and device. Background technique [0002] Channel modeling is an important research direction in wireless communication, and an accurate channel model is a prerequisite for any wireless communication system design and performance analysis. The main purpose of channel modeling is to accurately describe the statistical distribution of scattering paths in different environments. Among the models describing the statistical distribution of scattering paths in wireless channels, the tapped delay line model is quite representative. This model describes the channel as formed by the superposition of a large number of scattering paths in the delay domain, and includes small-scale fading characteristics. The TDL (Tapped Delay Line, tapped delay line) model has been widely used for a long time, and was adopted as a standard...

AI Technical Summary

Problems solved by technology

While many validation metrics exist to estimate the number of clusters, no metric can always correctly predict the number of clusters of scattering paths

Most studies still use visual recognition to get the optimal number of clusters in the environment, which greatly reduces the efficiency of automatic clustering algorithms

[0010] 3) Most clustering algorithms still require many preset parameters manually entered

For example, in the KPM algorithm, the initial information of the cluster (time delay and angle) needs to be defined, and the weight parameters of time delay and angle also need to be adjusted repeatedly to obtain a reasonable initialization parameter itself is also very difficult

Images